Field of Science

And with that I'll take both good and bad questions...

When scientists give talks/seminars/colloquia/presentations/speeches/lectures/lessons/defenses/instructions, questions are welcomed, and (probably to give themselves time to think), they often start by stating that that is a good question.

What they don't do is designate a question as dumb or stupid, because that's offensive, and scientists are frail like quails when it comes to their intellects.

"Good" questions:
  1. I've never thought of that before, but if you'll allow me to speculate...
  2. I have thought about that before, but I haven't come up with an answer, and here's why.
  3. I have thought about that before, and here's the answer.
"Dumb" questions
  1. I have already explained that, but I'm happy to go over it again.
  2. The answer is obvious, and I'm surprised you don't know it.
I welcome all of these myself, but there are also questions that can be designated "mean", which I enjoy less. For example, when the person asking already knows the question and is just quizzing you (acceptable at a defense/exam). Or when asked in order to bolster oneself showing off how much they know, or to turn the discussion in a certain direction (the aging Prof's pet area of study, typically), which is mean to the extent that it takes the focus off the speaker and the subject matter.

Scientists, especially graduate students, are very afraid of asking dumb questions, but they shouldn't be. First of all, if you ask one of the two kinds of dumb questions listed here, know that if you didn't get it, most likely you're not alone, and your peers will be grateful that you asked the question. On top of that, the speaker will often realize that maybe they shouldn't have assumed that everyone knew the answer, and should have been more careful explaining it in the first place. But also, I have found that learning to embrace what might by some be seen as a stupid question crucially paves the way to understanding. The road to failure is paved with unasked questions, or something like that.

Step Acuity

Please help me with a little science experiment.

Measure your Step Acuity (SA) and report it here.

Your SA is the number of steps before you hit a mark that you know whether you will hit the mark with your right or left foot. As you are walking down the street, choose a mark (a line or a spot) on the ground far in front of you, and say "right" or "left" out loud the moment you realize which foot is going to step on the mark, and from there count 1, 2, 3... until you hit the mark. Do you have an SA (i.e., mostly the same outcome every time), or is the number of steps highly variable?

No cheating, this is not a competition. Just honestly report back, SVP.

The hypothesis is that everyone has an SA, meaning the number of steps is highly invariable, and that SA is normally distributed. H2: The SA is different between those who played football (soccer) as a child and those who did not.

Also, feel free to report how many steps before hitting the mark you know that you are not hitting it at all (i.e., it's right in between left and right foot). Have fun.


The Ivory Power Manifesto

The Ivory Power Manifesto

The Ivory Tower is not my friend. We will not sit idle and look down at the masses while society goes to hell. Science is to be shared, and that's why we do it.

There is no such thing as "their terms". We bring our terms to the table always, and this is how we move forward.

When someone attacks, we defend ourselves. We won't just let them walk into our living room and destroy everything.

Religion leads to creationism. Nothing else does. Creationism thus implies religion. If you're still in doubt what the real problem here is, then you're part of the problem.

This is a war that will take a long time, but we know who will win it. We will not stand idle by and be on the wrong side of history.

We will use any and every opportunity to share science with the public.

The evidence can take us where it wants. We shall have no baggage that will make us resist it.

On some topics there is no legitimate scientific debate, but we will not be afraid of debating science-deniers on these topics, even though the public may perceive that as an acknowledgement that a legitimate debate exists.

Stating that there is no debate is a case of Expert Blindness. Just because you have resolved the issue doesn't mean everyone else has. We will share our expert knowledge with everyone.

Evolutionary forecasting and the tape metaphor

You may not think so (well, not you!), but evolutionary biology does make predictions. Particularly about the past, for example about the existence of moths with long proboscises to match an orchid's long nectary, or that no remains of rabbits will ever be found in the precambrian. But people generally don't talk about predicting future evolutionary outcomes, and those that do most often say that it is impossible - often referring to SJ Gould's metaphor of rewinding the tape of life (more on that below)

One could very well say the same thing about the weather: a hundred years ago, no one in their right mind (which is a tricky way of denoting all but the ones with the most foresight) would have guessed that we would ever be able to accurately predict the weather ten days into the future. But incredible advances in physics now allows us to do just that. Even though the weather systems are very sensitive to many factors, with the combination of the Navier-Stokes equations and a continuous effort to measure weather all over the world, we are now comfortable trusting that the weather forecasts are more accurate than guessing.

In this article by Carl Zimmer from 2014 several researchers explain how the same type of forecasting could potentially be applied to evolutionary biology. Particularly, it is not at all inconceivable that we will be able to predict flu virus evolution, thereby better making vaccines and saving lives. Microbial evolution, it turns out, often has much more repeatable outcomes than what is traditionally thought. Recent research strongly suggest that this is the case, for example when bacteria evolve the same solutions again and again to the same problems.

Here's a simulation with an example of that:



In this very simple example, you'll have to admit it's pretty easy to predict where the population is going to go, yes? (See the full video on youtube for more about fitness landscapes.)

One way to think of evolutionary forecasting is by obtaining as much data as possible about the fitness of genetic variants. If we know which variants (bacteria and viruses are the best candidate organisms) are most successful in making copies of themselves, then we can potentially make predictions of how they will evolve in the near future - just like with the weather.

The problem is that measuring fitness for many many genetic variants is still difficult. It means sequencing many individual organisms, and it means measuring fitness for each one of those variants. The resulting fitness landscape (i.e., fitness as a function of the genotype) is the last of three parameters needed to predict the simplest evolutionary systems. The other two are the population size and the mutation rate. Both of these are fortunately much easier to measure than the fitness landscape. So the fitness landscape remains the biggest obstacle. But it has only taken a few years to get quite far with measurements, and I don’t think it is at all inconceivable that we could have evolutionary forecasts for certain viruses and bacteria within the next few decades.

Returning to Gould, I predict that his lasting impact on evolutionary biology will be similar to Lamarck's: His "rewinding the tape of life" will be the quintessential error that people will keep referring to when talking about evolutionary forecasting.

To be more specific: I predict that 187 years from Gould's death (i.e., by 2189), when his name is mentioned, most people will immediately think of the tape first. And additionally that the idea has been somewhat refuted, since they then will be pretty good at evolutionary forecasting. (Lamarck died 187 years ago.)

My friend Larry Moran, who blogs at Sandwalk, is a staunch Gould admirer, and he asked me these questions in regard to my prediction:
Do you really believe that if we reran the tape of life we would always end up with a continent like Australia where marsupials outnumber placental mammals? 
Do you believe that every possible scenario would result in hundreds of species of dinosaurs that go extinct about 65 million years ago? 
Is the great Permian extinction always going to happen? 
Will an oxygen evolving complex always arise in primitive cyanobacterium about 2.5 billion years ago? 
Are you convinced that replaying the tape of life will always produce eukaryotes with introns and complex spliceosomes? 
Will no replay ever produce highly intelligent New World Monkeys? 
Are poisonous mushrooms absolutely necessary? 
Will the platypus always evolve and never go extinct? 
If we ever find another planet that's very similar to Earth do you expect to find intelligent mammals and maple trees?
And Larry made his own prediction:
Gould will be remembered for his attack on adaptationism and reminding us that there's more to evolution than natural selection. He will be remembered for teaching us about contingency and exaptation. He will be remembered for hierarchical theory. Above all, he will be remembered for making us all aware of the fact that evolution is a lot more complicated than we think.
One objection to the tape metaphor is that it is not well-defined. Does "rewinding the tape of life" mean that everything starts out the same? The initial conditions are *exactly* the same? What about the random events that influence evolution, will they be exactly the same or are they different? Are we talking about the environmental effects (weather, plate tectonics, ET events) being the same or not? Would mutations be identical or just drawn from the same distribution? Some people may have a clear idea of these things and thus feel confident of what the metaphor really means, but I know for a fact that there is not a general consensus about this, with several evolutionary biologists that I have talked to having differing views about its interpretation. We can all agree that randomness is very important in evolution, but not what kind of randomness is part of the metaphor.

In answer to Larry's questions above, one answer is no, not that specifically. Stochasticity (randomness) is too strong a factor, and future evolutionary events are contingent on earlier events.

On the other hand, it may be that stochasticity doesn't mean that evolution is unpredictable, just like it is not unpredictable what will happen when you heat a pot of water, even though we can't predict the path of every individual water molecule.

On the first hand again, if we're talking about, say, evolution on other planets, then no, we cannot expect those exact outcomes. However, Gould also denied that humans would be a repeat outcome if the tape was rewinded, but it is not clearly defined what we mean by "humans" in this context. If by humans we mean intelligent bipedal tetrapods with a head on top, then I do think there are very good reasons to believe that such creatures could evolve again and again, and be a feature on many planets that harbor life throughout the universe.

Back on the second hand, one can argue that every event that ever happens is determined by previous physical states. Things are only as random as we are unable to foresee the next event. A random mutation is only random because we don't know the factors that determine it, or because we don't have the data to predict it. In other words, if we rewind until some time point, everything else is the same up until that time point, so there are good reasons to believe that all future random events would be identical to the first run, and we would thus get exactly what we got the first time.